人工智慧面試之回應性對應徵者印象管理之影響

Abstract

隨著網路世代進步以及智慧型裝置普及化，企業為了有效提升甄選的效率，將人工智慧應用於錄影面試情境中，透過分析應徵者之微表情來判讀其人格特質與溝通技巧。然而，應徵者是否會因人工智慧錄影面試中而感受到焦慮並出現面試表現焦慮的狀況是未知的研究議題。甚者，不同介面設計的人工智慧錄影面試是否會影響應徵者的印象管理行為，包括誠實與欺騙式的，更是一個待研究的議題。本研究基於社交介面理論與欺騙可能性模型，在錄影面試系統設計回應性介面，運用了兩種不同的人工智慧錄影面試介面進行實驗(是否具備回應性)(N=64)，面試過程中透過聲紋與回應語讓應徵者感受到回應性，藉此期望能夠影響應徵者印象管理行為研究結果發現，回應性介面會對於應徵者誠實型印象管理行為並無顯著影響，卻會提升欺騙型印象管理行為，另外卻能降低應徵者的面試表現焦慮。推論比起社交介面理論，遊戲化情境與人機互動的角色地位更適用於人工智慧面試情境，此情境更容易激發應徵者的競爭驅力而促使應徵者展示不實回應。最後依據研究結果，提出理論貢獻與實務意涵，幫助人力資源從業者與系統商在設計甄選環境時有足夠的參考依據。

The advancement of the Internet generation and the popularization of smart devices. In order to effectively improve the efficiency of selections, companies apply Artificial Intelligence (AI) to video interview. The AI-based video interview can analyze the micro expressions of applicants to infer their personalities and communication skills. However, whether the automatic video interview affect job candidates’ performance anxiety is an unknown research question. Moreover, whether different interface design can influence job candidates’ honest and deceptive impressions management (IM) behavior has not been explored. . Based on social interface theory and faking likelihood model, this research designs immediacy interface in the video interview system, and uses two different AI video interview interfaces to conduct experiments (whether have immediacy or not) (N=64). Candidates feel immediacy through voiceprint and feedback on interface, which is expected to influence candidates’ impression management behaviors.The results of the study found that the immediacy interface had no significant effect on candidates' honest IM behavior of the candidates, but increased the deceptive IM behavior, while reduced the interview performance anxiety of the candidates. The study found that gamification and human-automation interaction may be more applicable for AI interview situations than social interface theory. Gamification more likely to stimulate candidates' competitive attitude and prompt candidates to show faking responses. Finally, based on the research results, the theoretical contributions and practical implications are proposed to help human resource and system vendors have sufficient reference when designing the selection tools.